SDMetrics
SDMetrics copied to clipboard
Create multi-table QualityReport
Problem Description
Create the multi table QualityReport class. Implement the generate method and functionality.
The generate method should take in the real and synthetic dataframes for computing the following:
Compute & saves breakdowns for all of the endorsed metrics —
- KSComplement
- TVComplement
- CorrelationSimilarity with both 'Pearson' and 'Spearman'
- ContingencySimilarity
- CardinalityShapeSimilarity
Use tqdm to communicate progress. After it's finished, it should print out the general score and a rough breakdown.
Expected behavior
>>> from sdmetrics.reports.multi_table import QualityReport
>>> report = QualityReport()
>>> report.generate(real_data, synthetic_data, metadata)
Creating report: 96%|█████████▌| 192/200 [01:21<00:03, 2.37it/s]
Overall Quality Score: 78%
Properties:
Column Shapes: 91%
Column Pair Trends: 74%
Table Relationships: 69% # available for multi table only